Reconstruct velocity field at cell centers

[cbegeman]

Ideally, Omega would include the capability to reconstruct fields from edges to cell centers. In the absence of this capability, we need to

1. Add mesh files to the cache with coeffs_reconstruct
2. Use mpas_tools reconstruction from normalVelocity to velocityZonal and velocityMeridional
   github.com/MPAS-Dev/MPAS-Tools/blob/297a1b0bf08c0d9e191c0b0e1d35ae315ee49617/conda_package/mpas_tools/vector/reconstruct.py#L2-L18

[cbegeman]

@xylar For (2), do you think we should construct the command line argument to vector_reconstruct or just use reconstruct_variable directly?

[xylar]

I think it's a little cleaner to use reconstruct_variable() directly but it doesn't matter too much.

[xylar]

Here's the Compass test case that we use to write out the coefficients for a given mesh by running MPAS-Ocean for one time step:
https://github.com/MPAS-Dev/compass/tree/main/compass/ocean/tests/global_ocean/files_for_e3sm/write_coeffs_reconstruct

[cbegeman]

@xylar I'll take this one on.

[cbegeman]

Update

Work has been started here #448 but files still need to be cached.

[cbegeman]

Addressed by #448